Reasoning and chat
Multi-turn conversation for analysis, drafting and planning, with answers returned as clean JSON.
IRIDIMAX is the AI system from Emporiant, open to developers through one API. Chat, code, retrieval, images and agents, running on privacy-first IRIDIUM infrastructure.
IRIDIMAX covers the work developers do every day, from the first prompt to a running agent, behind a single key.
Multi-turn conversation for analysis, drafting and planning, with answers returned as clean JSON.
Write, explain and review code in common languages, from your editor, scripts or CI.
Ground answers in your own documents with embeddings and search over your data.
Create visuals through the ComfyUI-backed image pipeline, called the same way as everything else.
Let the model call your functions and chain steps into workflows that run on their own.
Built on Emporiant's IRIDIUM platform for GDPR-conscious teams, with a local-first architecture.
Send JSON, get JSON. If your stack can make an HTTPS request, it can talk to IRIDIMAX.
/v1/chatConversation and code generation/v1/embeddingsVectors for search and retrieval/v1/imagesImage generation/v1/agents/runTool-calling agent runs/v1/modelsModels available to your keycurl https://api.iridimax.com/v1/chat \
-H "Authorization: Bearer $IRIDIMAX_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "iridimax-core",
"messages": [
{"role": "user", "content": "Summarise this changelog in three bullets."}
]
}'
import os, requests
r = requests.post(
"https://api.iridimax.com/v1/chat",
headers={"Authorization": f"Bearer {os.environ['IRIDIMAX_API_KEY']}"},
json={
"model": "iridimax-core",
"messages": [
{"role": "user", "content": "Summarise this changelog in three bullets."}
],
},
timeout=60,
)
print(r.json()["message"]["content"])
const res = await fetch("https://api.iridimax.com/v1/chat", {
method: "POST",
headers: {
"Authorization": `Bearer ${process.env.IRIDIMAX_API_KEY}`,
"Content-Type": "application/json",
},
body: JSON.stringify({
model: "iridimax-core",
messages: [
{ role: "user", content: "Summarise this changelog in three bullets." },
],
}),
});
const data = await res.json();
console.log(data.message.content);
Tell us what you are building. You receive an API key for Open Access.
Post to /v1/chat with your key in the Authorization header.
Move to higher limits, or to your own IRIDIUM ONE hardware, when the project needs it.
Pick the level that matches where your project is today. You can change it later.
Every request passes the same path, so access control and scaling stay in one place.
Any HTTPS client with your API key.
Authentication, rate limits and request filtering.
Distributes load across available IRIDIMAX nodes.
Models running locally on AMD hardware.
Clean results ready for your product.
IRIDIMAX is the AI system from Emporiant, a company based in Graz, Austria. It gives developers access to chat, code, retrieval, image and agent capabilities through one API.
You can request an API key and start building without a sales call. Limits depend on your tier, and you can move to higher limits when your project grows.
Any language that can send HTTPS requests works today. The examples above show curl, Python and JavaScript.
IRIDIMAX is built on Emporiant's privacy-first IRIDIUM platform. Contact us for details on data handling for your use case.
Yes. IRIDIUM ONE is Emporiant's on-premise AI appliance. You can add more boxes as demand grows, and load balancing spreads requests across them.
Send an email to office@emporiant.com and tell us what you are building. We will set you up with access.
Tell us what you are building and we will set you up with Open Access to IRIDIMAX.
office@emporiant.com